bike_rental - Datenbasis SQL

Der vorliegende Datensatz enthält Daten von der Vermietung von Fahrrädern in der Region San Francisco.

Die Datenbasis ist eine SQLite3-tabelle.

Der Datensatz hat viel Potential von Analysen. Zu Beginn soll es allerdings reichen, dass du deine bisherigen Fähigkeiten und Kenntnisse übst und vertiefst.

Aufgabe/Ziel:

         - mache dich mit den Daten vertraut
         - schließe dich an die Datenbank an und erkunde die Metadaten(welche Tabellen mit welchen Spalten 
           liegen vor? etc...
         - Welche Erkenntnisse kannst du ziehen. Visualisiere diese gern mit ein paar Plots. Es ist egal welche                    Bibliothek du verwendest

Du kannst erkunden was du möchtest. Ziel ist, dass du erkennst, was du bereits alles kannst!

Wofür stehen die Spalten?

Tabelle: Status

-station_id: station ID number

-bikes_available: number of available bikes

-docks_available: number of available docks

-time: date and time, PST

Tabelle: Station Information

-station_id: station ID number (corresponds to "station_id")

-name: name of station

-lat: latitude

-long: longitude

-dockcount: number of total docks at station

-landmark: city (San Francisco, Redwood City, Palo Alto, Mountain View, San Jose)

-installation: original date that station was installed. If station was moved, it is noted below.

Tabelle: TRIP DATA

-Trip ID: numeric ID of bike trip

-Duration: time of trip in seconds

-Start Date: start date of trip with date and time, in PST

-Start Station: station name of start station

-Start Terminal: numeric reference for start station

-End Date: end date of trip with date and time, in PST

-End Station: station name for end station

-End Terminal: numeric reference for end station

-Bike #: ID of bike used

-Subscription Type: Subscriber = annual or 30-day member; Customer = 24-hour or 3-day member

-Zip Code: Home zip code of subscriber (customers can choose to manually enter zip at kiosk however data is unreliable)

Tabelle: WEATHER DATA

Daily weather information per service area, provided from Weather Underground in PST. Weather is listed from north to south (San Francisco, Redwood City, Palo Alto, Mountain View, San Jose).

# benötigte Bibliotheken einlesen

import sqlite3
import pandas as pd
import numpy as np

# Verbindung zur Datenbank aufbauen

conn = sqlite3.connect('rental_bike.sqlite')
 

Tables Info

# Abfrage


my_query = """
            select name
            from sqlite_master
            where type = 'table'

"""

res_df = pd.read_sql(my_query, conn)
res_df

	name
0	station
1	status
2	trip
3	weather

Tablle 'STATION'

my_query = """
            
            PRAGMA table_info(station)

"""

res_df = pd.read_sql(my_query, conn)
res_df

	cid	name	type	notnull	dflt_value	pk
0	0	id	INTEGER	0	None	1
1	1	name	TEXT	0	None	0
2	2	lat	NUMERIC	0	None	0
3	3	long	NUMERIC	0	None	0
4	4	dock_count	INTEGER	0	None	0
5	5	city	TEXT	0	None	0
6	6	installation_date	TEXT	0	None	0

!!! Achtung !!!: Spalte 'installation_date' ist im Textformat.

my_query = """
            
            select count(1)
            from station

"""

res_df = pd.read_sql(my_query, conn)
res_df

	count(1)
0	70

my_query = """
            
            select city, count(1)
            from station
            group by city

"""

res_df = pd.read_sql(my_query, conn)
res_df

	city	count(1)
0	Mountain View	7
1	Palo Alto	5
2	Redwood City	7
3	San Francisco	35
4	San Jose	16

-- Tabele 'STATION' enthält die Information über die 70 Stationen aus 5 Städten in California.

my_query = """
            
            select *
            from station

"""
# limit 10

station_info_df = pd.read_sql(my_query, conn)
station_info_df.head()

	id	name	lat	long	dock_count	city	installation_date
0	2	San Jose Diridon Caltrain Station	37.329732	-121.901782	27	San Jose	8/6/2013
1	3	San Jose Civic Center	37.330698	-121.888979	15	San Jose	8/5/2013
2	4	Santa Clara at Almaden	37.333988	-121.894902	11	San Jose	8/6/2013
3	5	Adobe on Almaden	37.331415	-121.893200	19	San Jose	8/5/2013
4	6	San Pedro Square	37.336721	-121.894074	15	San Jose	8/7/2013

Tablle 'STATUS'

my_query = """
            
            PRAGMA table_info(status)

"""

res_df = pd.read_sql(my_query, conn)
res_df

	cid	name	type	notnull	dflt_value	pk
0	0	station_id	INTEGER	0	None	0
1	1	bikes_available	INTEGER	0	None	0
2	2	docks_available	INTEGER	0	None	0
3	3	time	TEXT	0	None	0

!!! Achtung !!!: Spalte 'time' ist im Textformat.

my_query = """
            
            select count(1)
            from status

"""

res_df = pd.read_sql(my_query, conn)
res_df

	count(1)
0	71984434

my_query = """
            
            select *
            from status
            limit 10

"""

res_df = pd.read_sql(my_query, conn)
res_df

	station_id	bikes_available	docks_available	time
0	2	2	25	2013/08/29 12:06:01
1	2	2	25	2013/08/29 12:07:01
2	2	2	25	2013/08/29 12:08:01
3	2	2	25	2013/08/29 12:09:01
4	2	2	25	2013/08/29 12:10:01
5	2	2	25	2013/08/29 12:11:01
6	2	2	25	2013/08/29 12:12:01
7	2	2	25	2013/08/29 12:13:01
8	2	2	25	2013/08/29 12:15:01
9	2	2	25	2013/08/29 12:16:02

Info über der Status der Stationen wurde jeder Minute in die Tabelle 'STATUS' gespeichert.

time fotmat: 2013/08/29 12:06:01 -> YYYY/MM/DD MM:HH:SS

my_query = """
            
            select distinct station_id
            from status

"""

res_df = pd.read_sql(my_query, conn)
res_df

	station_id
0	2
1	3
2	4
3	5
4	6
...	...
65	77
66	80
67	82
68	83
69	84

70 rows × 1 columns

my_query = """
            
            select max(time)
            from status

"""

res_df = pd.read_sql(my_query, conn)
res_df

	max(time)
0	2015-08-31 23:59:02

my_query = """
            
            select min(time)
            from status

"""

res_df = pd.read_sql(my_query, conn)
res_df

	min(time)
0	2013/08/29 12:06:01

-- Die Information in der Tabelle 'STATUS' verfügt von 2013/08/29 12:06:01 bis zum 2015-08-31 23:59:02

# my_query = """
            
#             select *
#             from status
#             where station_id in (select id from station where city = 'San Francisco') 
#             and substr(time, 1, 4)||substr(time, 6, 2)||substr(time, 9, 2)||substr(time, 12, 2)||substr(time, 15, 2) >= '201404140000'
#             and substr(time, 1, 4)||substr(time, 6, 2)||substr(time, 9, 2)||substr(time, 12, 2)||substr(time, 15, 2) <= '201404202359'

# """

# res_df = pd.read_sql(my_query, conn)
# res_df

            
            
# #             and time >= '2014/04/14 00:00:00'
# #             and time <= '2014/04/19 23:59:59'

# df_sf_14_19_april_2014 = res_df.copy()

my_query = """
            
            select *
            from status
            where station_id = 70

"""

station_70_df = pd.read_sql(my_query, conn)
station_70_df

	station_id	bikes_available	docks_available	time
0	70	0	19	2013/08/29 12:06:01
1	70	0	19	2013/08/29 12:07:01
2	70	0	19	2013/08/29 12:08:01
3	70	0	19	2013/08/29 12:09:01
4	70	0	19	2013/08/29 12:10:01
...	...	...	...	...
1047135	70	15	4	2015-08-31 23:55:02
1047136	70	15	4	2015-08-31 23:56:01
1047137	70	15	4	2015-08-31 23:57:02
1047138	70	15	4	2015-08-31 23:58:02
1047139	70	15	4	2015-08-31 23:59:02

1047140 rows × 4 columns

my_query = """
            
            select *
            from status
            where bikes_available = 0 or docks_available = 0

"""

no_bikes_or_docks_df = pd.read_sql(my_query, conn)
no_bikes_or_docks_df

	station_id	bikes_available	docks_available	time
0	2	0	27	2013/11/03 02:00:01
1	2	0	27	2013/11/03 02:01:01
2	2	0	27	2013/11/03 02:02:01
3	2	0	27	2013/11/03 02:03:01
4	2	0	27	2013/12/04 10:02:03
...	...	...	...	...
855438	84	0	15	2015-08-27 16:26:02
855439	84	0	15	2015-08-27 16:27:02
855440	84	0	15	2015-08-27 16:28:02
855441	84	0	15	2015-08-27 16:29:02
855442	84	0	15	2015-08-27 16:30:02

855443 rows × 4 columns

my_query = """
            
            select *
            from status
            where bikes_available = 1 or docks_available = 1

"""

critical_bikes_or_docks_df = pd.read_sql(my_query, conn)
critical_bikes_or_docks_df

	station_id	bikes_available	docks_available	time
0	2	1	26	2013/08/29 20:38:01
1	2	1	26	2013/08/29 20:39:01
2	2	1	26	2013/08/29 20:40:02
3	2	1	26	2013/08/29 20:41:01
4	2	1	26	2013/08/29 20:42:01
...	...	...	...	...
1755764	84	1	14	2015-08-27 13:02:02
1755765	84	1	14	2015-08-27 13:03:02
1755766	84	1	14	2015-08-27 13:04:02
1755767	84	1	14	2015-08-27 13:05:02
1755768	84	1	14	2015-08-27 16:31:02

1755769 rows × 4 columns

my_query = """
            
            select *
            from status
            where station_id in (select id from station where city = 'San Francisco') 
            and substr(time, 1, 4)||substr(time, 6, 2)||substr(time, 9, 2)||substr(time, 12, 2)||substr(time, 15, 2) >= '201404160000'
            and substr(time, 1, 4)||substr(time, 6, 2)||substr(time, 9, 2)||substr(time, 12, 2)||substr(time, 15, 2) <= '201404162359'

"""

df_sf_16_april_2014 = pd.read_sql(my_query, conn)
df_sf_16_april_2014

	station_id	bikes_available	docks_available	time
0	39	14	5	2014-04-16 00:00:03
1	39	14	5	2014-04-16 00:01:03
2	39	14	5	2014-04-16 00:02:03
3	39	14	5	2014-04-16 00:03:02
4	39	14	5	2014-04-16 00:04:02
...	...	...	...	...
50395	82	6	9	2014-04-16 23:55:03
50396	82	6	9	2014-04-16 23:56:02
50397	82	6	9	2014-04-16 23:57:03
50398	82	6	9	2014-04-16 23:58:03
50399	82	6	9	2014-04-16 23:59:03

50400 rows × 4 columns

my_query = """ SELECT 
                station_id,
                avg(bikes_available) as avg_bikes_available,
                avg(docks_available) as avg_docks_available,
                strftime('%w',substr(time, 1, 4)||'-'||substr(time, 6, 2)||'-'||substr(time, 9, 2)||' '||substr(time, 12, 2)||':'||substr(time, 15, 2)||':'||'00') as day_of_week,
                substr(time, 12, 2) as h
                 
                FROM status
                GROUP BY station_id, day_of_week, h
             """

station_status_df = pd.read_sql(my_query,con=conn)
station_status_df

	station_id	avg_bikes_available	avg_docks_available	day_of_week	h
0	2	13.675880	13.295195	0	00
1	2	13.610708	13.360264	0	01
2	2	13.417803	13.552636	0	02
3	2	13.505408	13.465537	0	03
4	2	13.496455	13.474541	0	04
...	...	...	...	...	...
11755	84	7.027626	7.958676	6	19
11756	84	7.094770	7.891528	6	20
11757	84	7.171005	7.815297	6	21
11758	84	7.242466	7.743836	6	22
11759	84	7.290183	7.696119	6	23

11760 rows × 5 columns

Tablle 'TRIP'

my_query = """
            
            PRAGMA table_info(trip)

"""

res_df = pd.read_sql(my_query, conn)
res_df

	cid	name	type	notnull	dflt_value	pk
0	0	id	INTEGER	0	None	1
1	1	duration	INTEGER	0	None	0
2	2	start_date	TEXT	0	None	0
3	3	start_station_name	TEXT	0	None	0
4	4	start_station_id	INTEGER	0	None	0
5	5	end_date	TEXT	0	None	0
6	6	end_station_name	TEXT	0	None	0
7	7	end_station_id	INTEGER	0	None	0
8	8	bike_id	INTEGER	0	None	0
9	9	subscription_type	TEXT	0	None	0
10	10	zip_code	INTEGER	0	None	0

time -> TEXT !!!!!!!!!!!!!!

my_query = """
            
            select count(1)
            from trip

"""

res_df = pd.read_sql(my_query, conn)
res_df

	count(1)
0	669959

-- Es gibt die Information fast über 700 Tsd. Reisen.

my_query = """
            
            select subscription_type, count(1)
            from trip
            group by subscription_type

"""

res_df = pd.read_sql(my_query, conn)
res_df

	subscription_type	count(1)
0	Customer	103213
1	Subscriber	566746

my_query = """
            
            select *
            from trip
            limit 10

"""

res_df = pd.read_sql(my_query, conn)
res_df

	id	duration	start_date	start_station_name	start_station_id	end_date	end_station_name	end_station_id	bike_id	subscription_type	zip_code
0	4069	174	8/29/2013 9:08	2nd at South Park	64	8/29/2013 9:11	2nd at South Park	64	288	Subscriber	94114
1	4073	1067	8/29/2013 9:24	South Van Ness at Market	66	8/29/2013 9:42	San Francisco Caltrain 2 (330 Townsend)	69	321	Subscriber	94703
2	4074	1131	8/29/2013 9:24	South Van Ness at Market	66	8/29/2013 9:43	San Francisco Caltrain 2 (330 Townsend)	69	317	Subscriber	94115
3	4075	1117	8/29/2013 9:24	South Van Ness at Market	66	8/29/2013 9:43	San Francisco Caltrain 2 (330 Townsend)	69	316	Subscriber	94122
4	4076	1118	8/29/2013 9:25	South Van Ness at Market	66	8/29/2013 9:43	San Francisco Caltrain 2 (330 Townsend)	69	322	Subscriber	94597
5	4078	3829	8/29/2013 9:31	Redwood City Caltrain Station	22	8/29/2013 10:34	Redwood City Caltrain Station	22	228	Customer	94062
6	4079	995	8/29/2013 9:35	South Van Ness at Market	66	8/29/2013 9:52	South Van Ness at Market	66	327	Subscriber	94102
7	4080	764	8/29/2013 9:36	South Van Ness at Market	66	8/29/2013 9:49	San Francisco Caltrain 2 (330 Townsend)	69	315	Subscriber	94117
8	4081	218	8/29/2013 9:38	Mountain View City Hall	27	8/29/2013 9:41	Mountain View City Hall	27	150	Subscriber	97214
9	4084	287	8/29/2013 9:41	Mountain View City Hall	27	8/29/2013 9:46	Mountain View City Hall	27	138	Customer	97214

my_query = """
            
            select *
            from trip

"""

trip_customer_df = pd.read_sql(my_query, conn)
trip_customer_df.head(5)

	id	duration	start_date	start_station_name	start_station_id	end_date	end_station_name	end_station_id	bike_id	subscription_type	zip_code
0	4069	174	8/29/2013 9:08	2nd at South Park	64	8/29/2013 9:11	2nd at South Park	64	288	Subscriber	94114
1	4073	1067	8/29/2013 9:24	South Van Ness at Market	66	8/29/2013 9:42	San Francisco Caltrain 2 (330 Townsend)	69	321	Subscriber	94703
2	4074	1131	8/29/2013 9:24	South Van Ness at Market	66	8/29/2013 9:43	San Francisco Caltrain 2 (330 Townsend)	69	317	Subscriber	94115
3	4075	1117	8/29/2013 9:24	South Van Ness at Market	66	8/29/2013 9:43	San Francisco Caltrain 2 (330 Townsend)	69	316	Subscriber	94122
4	4076	1118	8/29/2013 9:25	South Van Ness at Market	66	8/29/2013 9:43	San Francisco Caltrain 2 (330 Townsend)	69	322	Subscriber	94597

trip_customer_df.shape

(669959, 11)

my_query = """
            
            select count(1)
            from
                (select distinct bike_id
                from trip)

"""

res_df = pd.read_sql(my_query, conn)
res_df

	count(1)
0	700

-- Es steht ein Bestand von 700 Fahrrädern zum Ausleihen zur Verfügung.

Tabelle 'WEATHER'

my_query = """
            
            PRAGMA table_info(weather)

"""

res_df = pd.read_sql(my_query, conn)
res_df

	cid	name	type	notnull	dflt_value	pk
0	0	date	TEXT	0	None	0
1	1	max_temperature_f	INTEGER	0	None	0
2	2	mean_temperature_f	INTEGER	0	None	0
3	3	min_temperature_f	INTEGER	0	None	0
4	4	max_dew_point_f	INTEGER	0	None	0
5	5	mean_dew_point_f	INTEGER	0	None	0
6	6	min_dew_point_f	INTEGER	0	None	0
7	7	max_humidity	INTEGER	0	None	0
8	8	mean_humidity	INTEGER	0	None	0
9	9	min_humidity	INTEGER	0	None	0
10	10	max_sea_level_pressure_inches	NUMERIC	0	None	0
11	11	mean_sea_level_pressure_inches	NUMERIC	0	None	0
12	12	min_sea_level_pressure_inches	NUMERIC	0	None	0
13	13	max_visibility_miles	INTEGER	0	None	0
14	14	mean_visibility_miles	INTEGER	0	None	0
15	15	min_visibility_miles	INTEGER	0	None	0
16	16	max_wind_Speed_mph	INTEGER	0	None	0
17	17	mean_wind_speed_mph	INTEGER	0	None	0
18	18	max_gust_speed_mph	INTEGER	0	None	0
19	19	precipitation_inches	INTEGER	0	None	0
20	20	cloud_cover	INTEGER	0	None	0
21	21	events	TEXT	0	None	0
22	22	wind_dir_degrees	INTEGER	0	None	0
23	23	zip_code	INTEGER	0	None	0

my_query = """
            
            select *
            from weather
            limit 10

"""

res_df = pd.read_sql(my_query, conn)
res_df

	date	max_temperature_f	mean_temperature_f	min_temperature_f	max_dew_point_f	mean_dew_point_f	min_dew_point_f	max_humidity	mean_humidity	min_humidity	...	mean_visibility_miles	min_visibility_miles	max_wind_Speed_mph	mean_wind_speed_mph	max_gust_speed_mph	precipitation_inches	cloud_cover	events	wind_dir_degrees	zip_code
0	8/29/2013	74	68	61	61	58	56	93	75	57	...	10	10	23	11	28	0	4		286	94107
1	8/30/2013	78	69	60	61	58	56	90	70	50	...	10	7	29	13	35	0	2		291	94107
2	8/31/2013	71	64	57	57	56	54	93	75	57	...	10	10	26	15	31	0	4		284	94107
3	9/1/2013	74	66	58	60	56	53	87	68	49	...	10	10	25	13	29	0	4		284	94107
4	9/2/2013	75	69	62	61	60	58	93	77	61	...	10	6	23	12	30	0	6		277	94107
5	9/3/2013	73	67	60	59	56	51	84	65	46	...	10	10	24	15	31	0	2		276	94107
6	9/4/2013	74	68	61	59	57	56	90	72	53	...	10	10	29	19	35	0	4		269	94107
7	9/5/2013	72	66	60	57	56	54	90	74	57	...	10	10	31	21	37	0	3		270	94107
8	9/6/2013	85	71	56	57	51	45	86	58	29	...	10	10	24	8	28	0	0		287	94107
9	9/7/2013	88	73	58	64	54	46	86	59	31	...	10	10	21	8	25	0	1		305	94107

10 rows × 24 columns

Data Frame Vorbereitung

# station_info_df
# trip_customer_df
# no_bikes_df
# full_station_df

station_info_df

station_info_df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 70 entries, 0 to 69
Data columns (total 7 columns):
 #   Column             Non-Null Count  Dtype  
---  ------             --------------  -----  
 0   id                 70 non-null     int64  
 1   name               70 non-null     object 
 2   lat                70 non-null     float64
 3   long               70 non-null     float64
 4   dock_count         70 non-null     int64  
 5   city               70 non-null     object 
 6   installation_date  70 non-null     object 
dtypes: float64(2), int64(2), object(3)
memory usage: 4.0+ KB

station_info_df.loc[:, ['city', 'lat', 'long']].head()

	city	lat	long
0	San Jose	37.329732	-121.901782
1	San Jose	37.330698	-121.888979
2	San Jose	37.333988	-121.894902
3	San Jose	37.331415	-121.893200
4	San Jose	37.336721	-121.894074

station_info_df.shape

(70, 7)

station_info_df.loc[:, 'id']

0      2
1      3
2      4
3      5
4      6
      ..
65    77
66    80
67    82
68    83
69    84
Name: id, Length: 70, dtype: int64

stations_trip = trip_customer_df.loc[:, 'start_station_id'].unique()

stations_trip

array([64, 66, 22, 27, 45, 35, 73, 75, 69, 76, 10, 49, 56, 70, 48, 67, 39,
       58, 61, 50, 47, 41, 74, 37, 42,  9, 55, 65, 77, 51, 68, 14, 34, 36,
       72, 11, 60, 62, 12, 46, 57,  5, 71,  6,  4, 28,  3,  8, 63,  7,  2,
       54, 59, 29, 13, 30, 23, 26, 21, 33, 38, 24, 16, 25, 80, 32, 31, 82,
       83, 84], dtype=int64)

len(stations_trip)

70

trip_customer_df

trip_customer_df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 669959 entries, 0 to 669958
Data columns (total 11 columns):
 #   Column              Non-Null Count   Dtype 
---  ------              --------------   ----- 
 0   id                  669959 non-null  int64 
 1   duration            669959 non-null  int64 
 2   start_date          669959 non-null  object
 3   start_station_name  669959 non-null  object
 4   start_station_id    669959 non-null  int64 
 5   end_date            669959 non-null  object
 6   end_station_name    669959 non-null  object
 7   end_station_id      669959 non-null  int64 
 8   bike_id             669959 non-null  int64 
 9   subscription_type   669959 non-null  object
 10  zip_code            669959 non-null  object
dtypes: int64(5), object(6)
memory usage: 56.2+ MB

trip_customer_df = trip_customer_df.merge(station_info_df.loc[:, ['id', 'city', 'lat', 'long']], left_on='start_station_id', right_on='id',
          suffixes=('', '_start'))

trip_customer_df.shape

(669959, 15)

trip_customer_df.drop(columns=['id_start'], inplace=True)

trip_customer_df.rename(columns={"city" : "city_start", 'lat': 'lat_start', 'long': 'long_start'}, inplace=True)

trip_customer_df.head(5)

	id	duration	start_date	start_station_name	start_station_id	end_date	end_station_name	end_station_id	bike_id	subscription_type	zip_code	city_start	lat_start	long_start
0	4069	174	8/29/2013 9:08	2nd at South Park	64	8/29/2013 9:11	2nd at South Park	64	288	Subscriber	94114	San Francisco	37.782259	-122.392738
1	4426	982	8/29/2013 12:52	2nd at South Park	64	8/29/2013 13:08	2nd at South Park	64	593	Subscriber	94117	San Francisco	37.782259	-122.392738
2	4465	419	8/29/2013 13:11	2nd at South Park	64	8/29/2013 13:18	Post at Kearney	47	498	Subscriber	94117	San Francisco	37.782259	-122.392738
3	4503	728	8/29/2013 13:29	2nd at South Park	64	8/29/2013 13:41	Embarcadero at Vallejo	48	331	Subscriber	94110	San Francisco	37.782259	-122.392738
4	4557	130	8/29/2013 13:57	2nd at South Park	64	8/29/2013 13:59	2nd at South Park	64	371	Subscriber	94122	San Francisco	37.782259	-122.392738

trip_customer_df = trip_customer_df.merge(station_info_df.loc[:, ['id', 'city', 'lat', 'long']], left_on='end_station_id', right_on='id',
          suffixes=('', '_end'))

trip_customer_df.drop(columns=['id_end'], inplace=True)

trip_customer_df.rename(columns={"city" : "city_end", 'lat': 'lat_end', 'long': 'long_end'}, inplace=True)

trip_customer_df.head(5)

	id	duration	start_date	start_station_name	start_station_id	end_date	end_station_name	end_station_id	bike_id	subscription_type	zip_code	city_start	lat_start	long_start	city_end	lat_end	long_end
0	4069	174	8/29/2013 9:08	2nd at South Park	64	8/29/2013 9:11	2nd at South Park	64	288	Subscriber	94114	San Francisco	37.782259	-122.392738	San Francisco	37.782259	-122.392738
1	4426	982	8/29/2013 12:52	2nd at South Park	64	8/29/2013 13:08	2nd at South Park	64	593	Subscriber	94117	San Francisco	37.782259	-122.392738	San Francisco	37.782259	-122.392738
2	4557	130	8/29/2013 13:57	2nd at South Park	64	8/29/2013 13:59	2nd at South Park	64	371	Subscriber	94122	San Francisco	37.782259	-122.392738	San Francisco	37.782259	-122.392738
3	4564	1693	8/29/2013 14:04	2nd at South Park	64	8/29/2013 14:32	2nd at South Park	64	272	Subscriber	94005	San Francisco	37.782259	-122.392738	San Francisco	37.782259	-122.392738
4	4566	1699	8/29/2013 14:04	2nd at South Park	64	8/29/2013 14:33	2nd at South Park	64	371	Subscriber	94122	San Francisco	37.782259	-122.392738	San Francisco	37.782259	-122.392738

trip_customer_df.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 669959 entries, 0 to 669958
Data columns (total 17 columns):
 #   Column              Non-Null Count   Dtype  
---  ------              --------------   -----  
 0   id                  669959 non-null  int64  
 1   duration            669959 non-null  int64  
 2   start_date          669959 non-null  object 
 3   start_station_name  669959 non-null  object 
 4   start_station_id    669959 non-null  int64  
 5   end_date            669959 non-null  object 
 6   end_station_name    669959 non-null  object 
 7   end_station_id      669959 non-null  int64  
 8   bike_id             669959 non-null  int64  
 9   subscription_type   669959 non-null  object 
 10  zip_code            669959 non-null  object 
 11  city_start          669959 non-null  object 
 12  lat_start           669959 non-null  float64
 13  long_start          669959 non-null  float64
 14  city_end            669959 non-null  object 
 15  lat_end             669959 non-null  float64
 16  long_end            669959 non-null  float64
dtypes: float64(4), int64(5), object(8)
memory usage: 92.0+ MB

trip_customer_df.isnull().sum()

id                    0
duration              0
start_date            0
start_station_name    0
start_station_id      0
end_date              0
end_station_name      0
end_station_id        0
bike_id               0
subscription_type     0
zip_code              0
city_start            0
lat_start             0
long_start            0
city_end              0
lat_end               0
long_end              0
dtype: int64

trip_customer_df.loc[:, ['start_date', 'end_date']].head(5)

	start_date	end_date
0	8/29/2013 9:08	8/29/2013 9:11
1	8/29/2013 12:52	8/29/2013 13:08
2	8/29/2013 13:57	8/29/2013 13:59
3	8/29/2013 14:04	8/29/2013 14:32
4	8/29/2013 14:04	8/29/2013 14:33

trip_customer_df.loc[:, 'start_date'] = pd.to_datetime(trip_customer_df.loc[:, 'start_date'])

trip_customer_df.loc[:, 'end_date'] = pd.to_datetime(trip_customer_df.loc[:, 'end_date'])

trip_customer_df.dtypes

id                             int64
duration                       int64
start_date            datetime64[ns]
start_station_name            object
start_station_id               int64
end_date              datetime64[ns]
end_station_name              object
end_station_id                 int64
bike_id                        int64
subscription_type             object
zip_code                      object
city_start                    object
lat_start                    float64
long_start                   float64
city_end                      object
lat_end                      float64
long_end                     float64
dtype: object

trip_customer_df.loc[:, ['start_date', 'end_date']].head(5)

	start_date	end_date
0	2013-08-29 09:08:00	2013-08-29 09:11:00
1	2013-08-29 12:52:00	2013-08-29 13:08:00
2	2013-08-29 13:57:00	2013-08-29 13:59:00
3	2013-08-29 14:04:00	2013-08-29 14:32:00
4	2013-08-29 14:04:00	2013-08-29 14:33:00

trip_customer_df.loc[:, 'subscription_type'] = trip_customer_df.loc[:, 'subscription_type'].astype("category")

trip_customer_df.dtypes

id                             int64
duration                       int64
start_date            datetime64[ns]
start_station_name            object
start_station_id               int64
end_date              datetime64[ns]
end_station_name              object
end_station_id                 int64
bike_id                        int64
subscription_type           category
zip_code                      object
city_start                    object
lat_start                    float64
long_start                   float64
city_end                      object
lat_end                      float64
long_end                     float64
dtype: object

trip_customer_df.loc[:,["id","duration","start_station_id", "end_station_id", "bike_id"]] = trip_customer_df.loc[:,["id","duration","start_station_id", "end_station_id", "bike_id"]].apply(pd.to_numeric, downcast="integer")

trip_customer_df.dtypes

id                             int32
duration                       int32
start_date            datetime64[ns]
start_station_name            object
start_station_id                int8
end_date              datetime64[ns]
end_station_name              object
end_station_id                  int8
bike_id                        int16
subscription_type           category
zip_code                      object
city_start                    object
lat_start                    float64
long_start                   float64
city_end                      object
lat_end                      float64
long_end                     float64
dtype: object

# trip_customer_df.loc[:, 'zip_code'].astype("int")

trip_customer_df.replace({"zip_code":''}, np.nan, inplace=True )

trip_customer_df.isnull().sum()

id                       0
duration                 0
start_date               0
start_station_name       0
start_station_id         0
end_date                 0
end_station_name         0
end_station_id           0
bike_id                  0
subscription_type        0
zip_code              6619
city_start               0
lat_start                0
long_start               0
city_end                 0
lat_end                  0
long_end                 0
dtype: int64

trip_customer_df.loc[:,["lat_start","long_start","lat_end", "long_end"]] = trip_customer_df.loc[:,["lat_start","long_start", "lat_end", "long_end"]].apply(pd.to_numeric, downcast="float")

trip_customer_df.dtypes

id                             int32
duration                       int32
start_date            datetime64[ns]
start_station_name            object
start_station_id                int8
end_date              datetime64[ns]
end_station_name              object
end_station_id                  int8
bike_id                        int16
subscription_type           category
zip_code                      object
city_start                    object
lat_start                    float32
long_start                   float32
city_end                      object
lat_end                      float32
long_end                     float32
dtype: object

trip_customer_df['Year_start'] = trip_customer_df.start_date.dt.year

trip_customer_df['Month_start'] = trip_customer_df.start_date.dt.month

trip_customer_df['Day_start'] = trip_customer_df.start_date.dt.day

trip_customer_df['Day_of_week_start'] = trip_customer_df.start_date.dt.day_name() 

trip_customer_df['Day_of_week_num_start'] = trip_customer_df.start_date.dt.day_of_week #The day of the week with Monday=0, Sunday=6.

trip_customer_df['Time_start'] = trip_customer_df.start_date.dt.time

#trip_customer_df['rental_duration_sec'] = (trip_customer_df.loc[:, 'end_date'] - trip_customer_df.loc[:, 'start_date']).dt.total_seconds()
trip_customer_df['rental_duration'] = trip_customer_df.loc[:, 'end_date'] - trip_customer_df.loc[:, 'start_date']

# trip_customer_df.drop(columns=['Day_end'], inplace=True)
# trip_customer_df.drop(columns=['rental_duration_sec'], inplace=True)

trip_customer_df.head()

	id	duration	start_date	start_station_name	start_station_id	end_date	end_station_name	end_station_id	bike_id	subscription_type	...	city_end	lat_end	long_end	Year_start	Month_start	Day_start	Day_of_week_start	Day_of_week_num_start	Time_start	rental_duration
0	4069	174	2013-08-29 09:08:00	2nd at South Park	64	2013-08-29 09:11:00	2nd at South Park	64	288	Subscriber	...	San Francisco	37.782261	-122.392738	2013	8	29	Thursday	3	09:08:00	0 days 00:03:00
1	4426	982	2013-08-29 12:52:00	2nd at South Park	64	2013-08-29 13:08:00	2nd at South Park	64	593	Subscriber	...	San Francisco	37.782261	-122.392738	2013	8	29	Thursday	3	12:52:00	0 days 00:16:00
2	4557	130	2013-08-29 13:57:00	2nd at South Park	64	2013-08-29 13:59:00	2nd at South Park	64	371	Subscriber	...	San Francisco	37.782261	-122.392738	2013	8	29	Thursday	3	13:57:00	0 days 00:02:00
3	4564	1693	2013-08-29 14:04:00	2nd at South Park	64	2013-08-29 14:32:00	2nd at South Park	64	272	Subscriber	...	San Francisco	37.782261	-122.392738	2013	8	29	Thursday	3	14:04:00	0 days 00:28:00
4	4566	1699	2013-08-29 14:04:00	2nd at South Park	64	2013-08-29 14:33:00	2nd at South Park	64	371	Subscriber	...	San Francisco	37.782261	-122.392738	2013	8	29	Thursday	3	14:04:00	0 days 00:29:00

5 rows × 24 columns

trip_customer_df.loc[trip_customer_df['duration'] == trip_customer_df['duration'].max(), :]

	id	duration	start_date	start_station_name	start_station_id	end_date	end_station_name	end_station_id	bike_id	subscription_type	...	city_end	lat_end	long_end	Year_start	Month_start	Day_start	Day_of_week_start	Day_of_week_num_start	Time_start	rental_duration
135291	568474	17270400	2014-12-06 21:59:00	South Van Ness at Market	66	2015-06-24 20:18:00	2nd at Folsom	62	535	Customer	...	San Francisco	37.785297	-122.396233	2014	12	6	Saturday	5	21:59:00	199 days 22:19:00

1 rows × 24 columns

# Fahrräder, die für mehr als einen Tag ausgeliehen wurden

trip_customer_df.loc[trip_customer_df['duration']  > 60 * 60 * 24, :].sort_values('duration')

	id	duration	start_date	start_station_name	start_station_id	end_date	end_station_name	end_station_id	bike_id	subscription_type	...	city_end	lat_end	long_end	Year_start	Month_start	Day_start	Day_of_week_start	Day_of_week_num_start	Time_start	rental_duration
627343	897491	86553	2015-08-19 19:58:00	Castro Street and El Camino Real	32	2015-08-20 20:01:00	Evelyn Park and Ride	30	120	Customer	...	Mountain View	37.390278	-122.066551	2015	8	19	Wednesday	2	19:58:00	1 days 00:03:00
611014	709268	86573	2015-04-02 10:18:00	Cowper at University	37	2015-04-03 10:20:00	Cowper at University	37	188	Customer	...	Palo Alto	37.448597	-122.159508	2015	4	2	Thursday	3	10:18:00	1 days 00:02:00
645219	84179	86609	2013-11-05 12:24:00	San Salvador at 1st	8	2013-11-06 12:28:00	San Salvador at 1st	8	194	Customer	...	San Jose	37.330166	-121.885834	2013	11	5	Tuesday	1	12:24:00	1 days 00:04:00
608480	222321	86668	2014-03-21 08:22:00	San Antonio Caltrain Station	29	2014-03-22 08:27:00	San Antonio Caltrain Station	29	662	Customer	...	Mountain View	37.406940	-122.106758	2014	3	21	Friday	4	08:22:00	1 days 00:05:00
665985	48791	86890	2013-10-07 12:29:00	San Jose Civic Center	3	2013-10-08 12:38:00	San Jose Civic Center	3	118	Customer	...	San Jose	37.330696	-121.888977	2013	10	7	Monday	0	12:29:00	1 days 00:09:00
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
612403	111309	722236	2013-11-30 13:29:00	University and Emerson	35	2013-12-08 22:06:00	University and Emerson	35	247	Customer	...	Palo Alto	37.444523	-122.163094	2013	11	30	Saturday	5	13:29:00	8 days 08:37:00
613014	841176	1133540	2015-07-10 10:35:00	University and Emerson	35	2015-07-23 13:27:00	University and Emerson	35	262	Customer	...	Palo Alto	37.444523	-122.163094	2015	7	10	Friday	4	10:35:00	13 days 02:52:00
630246	750192	1852590	2015-05-02 06:17:00	San Antonio Shopping Center	31	2015-05-23 16:53:00	Castro Street and El Camino Real	32	680	Subscriber	...	Mountain View	37.385956	-122.083679	2015	5	2	Saturday	5	06:17:00	21 days 10:36:00
416121	825850	2137000	2015-06-28 21:50:00	Market at Sansome	77	2015-07-23 15:27:00	Yerba Buena Center of the Arts (3rd @ Howard)	68	466	Customer	...	San Francisco	37.784878	-122.401016	2015	6	28	Sunday	6	21:50:00	24 days 17:37:00
135291	568474	17270400	2014-12-06 21:59:00	South Van Ness at Market	66	2015-06-24 20:18:00	2nd at Folsom	62	535	Customer	...	San Francisco	37.785297	-122.396233	2014	12	6	Saturday	5	21:59:00	199 days 22:19:00

296 rows × 24 columns

# kategorien = ['Mon.','Tue.','Wed.', 'Thu.', 'Fri.', 'Sat.', 'Sun.']
# trip_customer_df["Day_of_week_start"] = pd.cut(trip_customer_df.loc[:, "Day_of_week_start"], bins = [-1, 0, 1, 2, 3,4,5,6] , labels = kategorien)
# trip_customer_df.head()

trip_customer_df.loc[:, 'subscription_type'].value_counts()

Subscriber    566746
Customer      103213
Name: subscription_type, dtype: int64

trip_customer_df.shape

(669959, 24)

trip_customer_df.to_pickle('trip_customer_df_after_preprocessing.pkl')

station_info_df.to_pickle('station_info_df.pkl')

trip_customer_df.duration.describe()

count    6.699590e+05
mean     1.107950e+03
std      2.225544e+04
min      6.000000e+01
25%      3.440000e+02
50%      5.170000e+02
75%      7.550000e+02
max      1.727040e+07
Name: duration, dtype: float64

Ohne Bikes oder Docks

no_bikes_or_docks_df.head()

	station_id	bikes_available	docks_available	time
0	2	0	27	2013/11/03 02:00:01
1	2	0	27	2013/11/03 02:01:01
2	2	0	27	2013/11/03 02:02:01
3	2	0	27	2013/11/03 02:03:01
4	2	0	27	2013/12/04 10:02:03

no_bikes_or_docks_df.loc[:, 'time'] = pd.to_datetime(no_bikes_or_docks_df.loc[:, 'time'])
no_bikes_or_docks_df.dtypes

station_id                  int64
bikes_available             int64
docks_available             int64
time               datetime64[ns]
dtype: object

no_bikes_df = no_bikes_or_docks_df.loc[no_bikes_or_docks_df.bikes_available == 0, :].groupby('station_id', as_index=False).size()
no_docks_df = no_bikes_or_docks_df.loc[no_bikes_or_docks_df.docks_available == 0, :].groupby('station_id', as_index=False).size()

station_problem_df = no_bikes_df.merge(no_docks_df, left_on='station_id', right_on='station_id', how='left')
station_problem_df.rename(columns={'size_x' : 'no_bikes_min', 'size_y' : 'no_docks_min'}, inplace=True)
station_problem_df.replace({"no_docks_min":np.nan}, 0 , inplace=True )
station_problem_df.loc[:, ["no_bikes_min","no_docks_min","station_id"]] = station_problem_df.loc[:,
                    ["no_bikes_min","no_docks_min","station_id"]].apply(pd.to_numeric, downcast="integer")
station_problem_df

	station_id	no_bikes_min	no_docks_min
0	2	1340	466
1	3	1066	4817
2	4	10618	6614
3	5	1321	0
4	6	3564	3633
...	...	...	...
65	77	4436	3129
66	80	2105	5
67	82	17824	3738
68	83	808	0
69	84	3642	1744

70 rows × 3 columns

station_problem_df.dtypes

station_id       int8
no_bikes_min    int16
no_docks_min    int16
dtype: object

Eine kritische Anzahl von Fahrrädern oder Docking-Stationen

critical_bikes_or_docks_df.head()

	station_id	bikes_available	docks_available	time
0	2	1	26	2013/08/29 20:38:01
1	2	1	26	2013/08/29 20:39:01
2	2	1	26	2013/08/29 20:40:02
3	2	1	26	2013/08/29 20:41:01
4	2	1	26	2013/08/29 20:42:01

critical_bikes_or_docks_df.loc[:, 'time'] = pd.to_datetime(critical_bikes_or_docks_df.loc[:, 'time'])
critical_bikes_or_docks_df.dtypes

station_id                  int64
bikes_available             int64
docks_available             int64
time               datetime64[ns]
dtype: object

SF Status 16.04.2014

df_sf_16_april_2014.head()

	station_id	bikes_available	docks_available	time
0	39	14	5	2014-04-16 00:00:03
1	39	14	5	2014-04-16 00:01:03
2	39	14	5	2014-04-16 00:02:03
3	39	14	5	2014-04-16 00:03:02
4	39	14	5	2014-04-16 00:04:02

df_sf_16_april_2014.dtypes

station_id          int64
bikes_available     int64
docks_available     int64
time               object
dtype: object

df_sf_16_april_2014.loc[:, 'time'] = pd.to_datetime(df_sf_16_april_2014.loc[:, 'time'])

df_sf_16_april_2014.loc[:, 
            ["station_id","bikes_available","docks_available"]] = df_sf_16_april_2014.loc[:,
                                                                        ["station_id","bikes_available","docks_available"]].apply(pd.to_numeric, downcast="integer")

df_sf_16_april_2014.dtypes

station_id                   int8
bikes_available              int8
docks_available              int8
time               datetime64[ns]
dtype: object

df_sf_16_april_2014 = df_sf_16_april_2014.merge(station_info_df.loc[:, ['id', 'lat', 'long']], left_on='station_id', right_on='id')
df_sf_16_april_2014.drop(columns=['id'], inplace=True)
df_sf_16_april_2014.head()

	station_id	bikes_available	docks_available	time	lat	long
0	39	14	5	2014-04-16 00:00:03	37.783871	-122.408433
1	39	14	5	2014-04-16 00:01:03	37.783871	-122.408433
2	39	14	5	2014-04-16 00:02:03	37.783871	-122.408433
3	39	14	5	2014-04-16 00:03:02	37.783871	-122.408433
4	39	14	5	2014-04-16 00:04:02	37.783871	-122.408433

df_sf_16_april_2014["time_of_day"] = df_sf_16_april_2014.time.dt.time
df_sf_16_april_2014["hour"] = df_sf_16_april_2014.time.dt.hour
df_sf_16_april_2014.head()

	station_id	bikes_available	docks_available	time	lat	long	time_of_day	hour
0	39	14	5	2014-04-16 00:00:03	37.783871	-122.408433	00:00:03	0
1	39	14	5	2014-04-16 00:01:03	37.783871	-122.408433	00:01:03	0
2	39	14	5	2014-04-16 00:02:03	37.783871	-122.408433	00:02:03	0
3	39	14	5	2014-04-16 00:03:02	37.783871	-122.408433	00:03:02	0
4	39	14	5	2014-04-16 00:04:02	37.783871	-122.408433	00:04:02	0

Visualisirung

Wo habe ich welche Stationen?

import plotly.express as px
import plotly.graph_objects as go

Verteilung der Fahrradverleihstationen nach Städten

city_station_df = station_info_df.groupby(['city'], as_index=False).size().sort_values(by='size')

city_station_df 

	city	size
1	Palo Alto	5
0	Mountain View	7
2	Redwood City	7
4	San Jose	16
3	San Francisco	35

fig = px.pie(city_station_df, values = 'size', names='city', title= 'Anzahl der Mietstationen in jeder Stadt')

fig.update_traces(textposition='inside', textinfo='percent+value')

fig.update_traces(texttemplate='%{percent}'+' '+'(%{value} St.)', textposition='inside')

fig.show()

-- Fahrradverleihstationen gibt es in fünf kalifornischen Städten

Fahrradverleihstationen auf der Karte

fig = px.scatter_mapbox(station_info_df,
                        lat="lat", lon="long",
                        color='dock_count',
                        mapbox_style="carto-positron",
                        size='dock_count',
                        size_max=15,
                        hover_data =["city","name","installation_date"],
                        title='Fahrradverleihstationen auf der Karte')
fig.show()

Verteilung der Fahrradverleihstationen nach Städten

city_station_df = station_info_df.groupby(['city'], as_index=False).size().sort_values(by='size')

city_station_df

	city	size
1	Palo Alto	5
0	Mountain View	7
2	Redwood City	7
4	San Jose	16
3	San Francisco	35

fig = px.pie(city_station_df, values = 'size', names='city', title= 'Anzahl der Stationen in jeder Stadt')

fig.update_traces(textposition='inside', textinfo='percent+value')

fig.update_traces(texttemplate='%{percent}'+' '+'(%{value} St.)', textposition='inside')

fig.show()

Top 10 der am häufigsten verwendeten Stationen

top_10_stationen_df = trip_customer_df.groupby(["start_station_id"], as_index=False).size().sort_values("size", ascending=False).iloc[0:10, :]
top_10_stationen_df = top_10_stationen_df.merge(station_info_df, left_on='start_station_id', right_on='id')
top_10_stationen_df.head()

	start_station_id	size	id	name	lat	long	dock_count	city	installation_date
0	70	49092	70	San Francisco Caltrain (Townsend at 4th)	37.776617	-122.395260	19	San Francisco	8/23/2013
1	69	33742	69	San Francisco Caltrain 2 (330 Townsend)	37.776600	-122.395470	23	San Francisco	8/23/2013
2	50	32934	50	Harry Bridges Plaza (Ferry Building)	37.795392	-122.394203	23	San Francisco	8/20/2013
3	60	27713	60	Embarcadero at Sansome	37.804770	-122.403234	15	San Francisco	8/21/2013
4	55	26089	55	Temporary Transbay Terminal (Howard at Beale)	37.789756	-122.394643	23	San Francisco	8/20/2013

fig = px.bar(top_10_stationen_df.sort_values("size"), x="size", y="name", orientation='h', color = "city",
            labels={'size':'Anzahl Fahrten', 'name':'Stationsname', 'city':'Stadt', 'id':'ID'},
            hover_data=['city', 'size','name', 'id'])
fig.show()

Daten der Installation der Stationen

station_info_df.loc[:,'installation_date'].unique()

array(['8/6/2013', '8/5/2013', '8/7/2013', '8/12/2013', '8/15/2013',
       '8/16/2013', '12/31/2013', '8/14/2013', '8/25/2013', '8/19/2013',
       '8/20/2013', '8/21/2013', '8/22/2013', '8/23/2013', '1/22/2014',
       '2/20/2014', '4/9/2014'], dtype=object)

station_info_df.loc[:,'installation_date'] = pd.to_datetime(station_info_df.loc[:,'installation_date'], format='%m/%d/%Y')

pd.to_datetime('8/6/2013', format='%m/%d/%Y')

Timestamp('2013-08-06 00:00:00')

tmp_df = station_info_df.groupby(['city', 'installation_date'], as_index=False).size()

city_inst_date_df = tmp_df.groupby(['city', 'installation_date']).sum().groupby(level=0).cumsum().reset_index().sort_values(by='installation_date')

city_inst_date_df.head()

	city	installation_date	size
15	San Jose	2013-08-05	5
16	San Jose	2013-08-06	10
17	San Jose	2013-08-07	14
5	Redwood City	2013-08-12	4
3	Palo Alto	2013-08-14	4

fig = px.line(city_inst_date_df, x="installation_date", y="size", color="city", markers=True,
             title='Daten der Installation der Stationen',
             labels={'size' : 'Gesamtzahl der Station in Stadt', 'city':'Stadt', 'installation_date':'Date der Installation'})
#fig.update_traces(textposition="bottom right")
fig.show()

Welche Customer habe ich?

df_subscription = trip_customer_df.groupby(['subscription_type'], as_index=False).size()

df_subscription 

	subscription_type	size
0	Customer	103213
1	Subscriber	566746

Die Gesamtzahl der Fahrten je nach Art des Abonnements

fig = px.bar(data_frame=df_subscription,x='subscription_type',y='size', text = 'size', title="The total number of trips depending on the type of subscription",
            labels={'size':'Anzahl der Fahrten', 'subscription_type':'Art des Abonnements'})
fig.update_traces(texttemplate='%{text}', textposition='outside')
fig.update_traces(showlegend=False)
# fig.update(layout_coloraxis_showscale=False)

---> Die meisten Benutzer haben ein Fahrradverleih-Abonnement

Die Gesamtzahl der Fahrten in Abhängigkeit vom Wochentag

df_days_trip = trip_customer_df.groupby(['Day_of_week_start', 'Day_of_week_num_start'], as_index=False).size().sort_values(by='Day_of_week_num_start')

df_days_trip 

	Day_of_week_start	Day_of_week_num_start	size
1	Monday	0	115873
5	Tuesday	1	122259
6	Wednesday	2	120201
4	Thursday	3	119089
0	Friday	4	109361
2	Saturday	5	44785
3	Sunday	6	38391

fig = px.bar(data_frame=df_days_trip,x='Day_of_week_start',y='size', text = 'size', title="Nutzung über die Wochentage",
            labels={'size':'Anzahl der Fahrten', 'Day_of_week_start':'Wochentag'})
fig.update_traces(texttemplate='%{text}', textposition='outside')
fig.update_traces(showlegend=False)

Gesamtzahl der Fahrten pro Wochentag je nach Abonnementstyp

df_days_trip = trip_customer_df.groupby(
    ['Day_of_week_start', 'Day_of_week_num_start', 'subscription_type'], as_index=False, observed=True
).size().sort_values(by=['Day_of_week_num_start', 'subscription_type'])

df_days_trip 

	Day_of_week_start	Day_of_week_num_start	subscription_type	size
3	Monday	0	Customer	11469
2	Monday	0	Subscriber	104404
11	Tuesday	1	Customer	11040
10	Tuesday	1	Subscriber	111219
13	Wednesday	2	Customer	11495
12	Wednesday	2	Subscriber	108706
9	Thursday	3	Customer	12451
8	Thursday	3	Subscriber	106638
1	Friday	4	Customer	14946
0	Friday	4	Subscriber	94415
5	Saturday	5	Customer	22125
4	Saturday	5	Subscriber	22660
7	Sunday	6	Customer	19687
6	Sunday	6	Subscriber	18704

fig = px.bar(df_days_trip, x='Day_of_week_start', y='size', title="Gesamtzahl der Fahrten je nach Wochentag und Abonnementtyp",
             text = 'size', color = 'subscription_type',
                         labels={'size':'Anzahl der Fahrten', 'Day_of_week_start':'Wochentag', 'subscription_type' : 'Abonnementstyp'})
fig.update_traces(texttemplate='%{text}', textposition='outside')

fig.show()

fig = px.line(df_days_trip, x="Day_of_week_start", y="size",
              title='Gesamtzahl der Fahrten je nach Wochentag und Abonnementtyp',
              color = 'subscription_type',
              labels={'size':'Anzahl der Fahrten', 'Day_of_week_start':'Wochentag', 'subscription_type' : 'Abonnementstyp'})
fig.update_traces(mode="markers+lines")

fig.show()

• Die Häufigkeit der Reisen für Kunden ohne Abonnement ("Subscriber") ist während der Woche nahezu gleich.
• Kunden mit einem 'Customer' Abonnement nutzen den Service nur an Wochentagen aktiv.

Gesamtzahl der Fahrer nach Stadt

df_city_trip = trip_customer_df.groupby(['city_start'], as_index=False).size().sort_values(by='size')

df_city_trip 

	city_start	size
2	Redwood City	3433
1	Palo Alto	6773
0	Mountain View	18167
4	San Jose	37878
3	San Francisco	603708

fig = px.pie(df_city_trip, values = 'size', names='city_start', title= 'Anzahl der Fahrten in jeder Stadt')

fig.show()

---> 90% der Reisen sind in San Francisco

Auswertung der Tripdaten

Zu welcher Tageszeit wird der Service am häufigsten genutzt?

trip_customer_df.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 669959 entries, 0 to 669958
Data columns (total 24 columns):
 #   Column                 Non-Null Count   Dtype          
---  ------                 --------------   -----          
 0   id                     669959 non-null  int32          
 1   duration               669959 non-null  int32          
 2   start_date             669959 non-null  datetime64[ns] 
 3   start_station_name     669959 non-null  object         
 4   start_station_id       669959 non-null  int8           
 5   end_date               669959 non-null  datetime64[ns] 
 6   end_station_name       669959 non-null  object         
 7   end_station_id         669959 non-null  int8           
 8   bike_id                669959 non-null  int16          
 9   subscription_type      669959 non-null  category       
 10  zip_code               663340 non-null  object         
 11  city_start             669959 non-null  object         
 12  lat_start              669959 non-null  float32        
 13  long_start             669959 non-null  float32        
 14  city_end               669959 non-null  object         
 15  lat_end                669959 non-null  float32        
 16  long_end               669959 non-null  float32        
 17  Year_start             669959 non-null  int64          
 18  Month_start            669959 non-null  int64          
 19  Day_start              669959 non-null  int64          
 20  Day_of_week_start      669959 non-null  object         
 21  Day_of_week_num_start  669959 non-null  int64          
 22  Time_start             669959 non-null  object         
 23  rental_duration        669959 non-null  timedelta64[ns]
dtypes: category(1), datetime64[ns](2), float32(4), int16(1), int32(2), int64(4), int8(2), object(7), timedelta64[ns](1)
memory usage: 95.2+ MB

trip_customer_df.head()

	id	duration	start_date	start_station_name	start_station_id	end_date	end_station_name	end_station_id	bike_id	subscription_type	...	city_end	lat_end	long_end	Year_start	Month_start	Day_start	Day_of_week_start	Day_of_week_num_start	Time_start	rental_duration
0	4069	174	2013-08-29 09:08:00	2nd at South Park	64	2013-08-29 09:11:00	2nd at South Park	64	288	Subscriber	...	San Francisco	37.782261	-122.392738	2013	8	29	Thursday	3	09:08:00	0 days 00:03:00
1	4426	982	2013-08-29 12:52:00	2nd at South Park	64	2013-08-29 13:08:00	2nd at South Park	64	593	Subscriber	...	San Francisco	37.782261	-122.392738	2013	8	29	Thursday	3	12:52:00	0 days 00:16:00
2	4557	130	2013-08-29 13:57:00	2nd at South Park	64	2013-08-29 13:59:00	2nd at South Park	64	371	Subscriber	...	San Francisco	37.782261	-122.392738	2013	8	29	Thursday	3	13:57:00	0 days 00:02:00
3	4564	1693	2013-08-29 14:04:00	2nd at South Park	64	2013-08-29 14:32:00	2nd at South Park	64	272	Subscriber	...	San Francisco	37.782261	-122.392738	2013	8	29	Thursday	3	14:04:00	0 days 00:28:00
4	4566	1699	2013-08-29 14:04:00	2nd at South Park	64	2013-08-29 14:33:00	2nd at South Park	64	371	Subscriber	...	San Francisco	37.782261	-122.392738	2013	8	29	Thursday	3	14:04:00	0 days 00:29:00

5 rows × 24 columns

trip_customer_df['Hour_start'] = trip_customer_df.start_date.dt.hour
trip_customer_df.head()

	id	duration	start_date	start_station_name	start_station_id	end_date	end_station_name	end_station_id	bike_id	subscription_type	...	lat_end	long_end	Year_start	Month_start	Day_start	Day_of_week_start	Day_of_week_num_start	Time_start	rental_duration	Hour_start
0	4069	174	2013-08-29 09:08:00	2nd at South Park	64	2013-08-29 09:11:00	2nd at South Park	64	288	Subscriber	...	37.782261	-122.392738	2013	8	29	Thursday	3	09:08:00	0 days 00:03:00	9
1	4426	982	2013-08-29 12:52:00	2nd at South Park	64	2013-08-29 13:08:00	2nd at South Park	64	593	Subscriber	...	37.782261	-122.392738	2013	8	29	Thursday	3	12:52:00	0 days 00:16:00	12
2	4557	130	2013-08-29 13:57:00	2nd at South Park	64	2013-08-29 13:59:00	2nd at South Park	64	371	Subscriber	...	37.782261	-122.392738	2013	8	29	Thursday	3	13:57:00	0 days 00:02:00	13
3	4564	1693	2013-08-29 14:04:00	2nd at South Park	64	2013-08-29 14:32:00	2nd at South Park	64	272	Subscriber	...	37.782261	-122.392738	2013	8	29	Thursday	3	14:04:00	0 days 00:28:00	14
4	4566	1699	2013-08-29 14:04:00	2nd at South Park	64	2013-08-29 14:33:00	2nd at South Park	64	371	Subscriber	...	37.782261	-122.392738	2013	8	29	Thursday	3	14:04:00	0 days 00:29:00	14

5 rows × 25 columns

df_days_hours = trip_customer_df.groupby(['Day_of_week_start', 'Day_of_week_num_start', 'Hour_start'], as_index=False).size().sort_values(by='Day_of_week_num_start')
df_days_hours

	Day_of_week_start	Day_of_week_num_start	Hour_start	size
47	Monday	0	23	515
27	Monday	0	3	49
28	Monday	0	4	182
29	Monday	0	5	593
30	Monday	0	6	2739
...	...	...	...	...
75	Sunday	6	3	82
74	Sunday	6	2	178
73	Sunday	6	1	291
82	Sunday	6	10	2728
83	Sunday	6	11	3402

168 rows × 4 columns

fig = px.scatter(df_days_hours, x="Day_of_week_start", y="Hour_start", size='size', size_max = 25, #text = 'size',
                labels={'Hour_start':'Uhrzeit der Fahrradabholung', 'Day_of_week_start':'Wochentag'},
                title = "Zu welchen Zeiten der Dienst am meisten genutzt wird")
# fig.update_traces(texttemplate='%{text}', textposition='middle right')

fig.show()

Das Diagramm zeigt, dass der Dienst an Wochentagen morgens und abends am meisten genutzt wird, an Wochenenden in der Mitte des Tages.

---> Stellen wir die durchschnittliche Wochenend- und Wochentagskurve dar.

df_days_hours.sort_values("Hour_start", inplace=True)
df_days_hours["Hour_start_str"] = df_days_hours.Hour_start.apply(lambda x: f"{x} Uhr")
df_days_hours["day_type"] = df_days_hours.Day_of_week_num_start.apply(lambda x: "Working" if x < 5  else "Weekend")
df_days_hours.head()

	Day_of_week_start	Day_of_week_num_start	Hour_start	size	Hour_start_str	day_type
72	Sunday	6	0	455	0 Uhr	Weekend
24	Monday	0	0	180	0 Uhr	Working
0	Friday	4	0	302	0 Uhr	Working
96	Thursday	3	0	277	0 Uhr	Working
144	Wednesday	2	0	253	0 Uhr	Working

df_day_type = df_days_hours.groupby(["day_type", "Hour_start", "Hour_start_str"], as_index=False).agg(size = ('size', np.mean))
df_day_type.sort_values("Hour_start", inplace=True)
df_day_type.head()

	day_type	Hour_start	Hour_start_str	size
0	Weekend	0	0 Uhr	469.0
24	Working	0	0 Uhr	246.6
25	Working	1	1 Uhr	116.6
1	Weekend	1	1 Uhr	303.0
2	Weekend	2	2 Uhr	177.0

fig = px.line_polar(df_day_type, r="size", theta="Hour_start_str", color="day_type", line_close=True,
            color_discrete_sequence=px.colors.qualitative.Plotly,
                   labels={'day_type':'Tag'})
fig.show()

Zu welcher Zeit Customer den Service nutzen, zu welcher Zeit Subscriber den Service nutzen

trip_customer_df['duration_h'] = trip_customer_df['duration'] / 3600
trip_customer_df["duration_min"] = trip_customer_df["duration"] / 60

customer_trips_df = trip_customer_df.loc[trip_customer_df.loc[:, 'subscription_type'] == 'Customer', :]
subscriber_trips_df = trip_customer_df.loc[trip_customer_df.loc[:, 'subscription_type'] == 'Subscriber', :]

df_days_hours_customer = customer_trips_df.groupby(
            ['Day_of_week_start', 'Day_of_week_num_start', 'Hour_start'],
            as_index=False).size().sort_values(by='Day_of_week_num_start')
df_days_hours_subscriber = subscriber_trips_df.groupby(
            ['Day_of_week_start', 'Day_of_week_num_start', 'Hour_start'],
            as_index=False).size().sort_values(by='Day_of_week_num_start')

fig = px.scatter(df_days_hours_customer, x="Day_of_week_start", y="Hour_start", size='size', size_max = 25, #text = 'size',
                labels={'Hour_start':'Uhrzeit der Fahrradabholung', 'Day_of_week_start':'Wochentag'},
                title = "Wie Customers den Service nutzen")

fig.show()

fig = px.scatter(df_days_hours_subscriber, x="Day_of_week_start", y="Hour_start", size='size', size_max = 25, #text = 'size',
                labels={'Hour_start':'Uhrzeit der Fahrradabholung', 'Day_of_week_start':'Wochentag'},
                title = "Wie Subscribers den Service nutzen")

fig.show()

Fahrtdauer

# station_tripcount_duration = trip_customer_df.groupby(["bike_id"], as_index=False).agg(
#             duration = ('duration_h', np.sum),
#             count = ('bike_id','size')).sort_values(by='bike_id', ascending=True)
# station_tripcount_duration

# station_tripcount_duration.describe()

# fig = px.scatter(
#     station_tripcount_duration.loc[station_tripcount_duration["duration"] < 1000, :],
#     x="count",
#     y="duration",
#     title="Nutzungsdauer (h) gegen Anzahl Nutzungen",
#     labels={'count':'Anzahl Fahrten', 'duration':'Gesamtreisedauer'},
#     hover_data=['count', 'duration','bike_id'])
# fig.show()

trip_customer_df.loc[:, ['duration', 'duration_min', 'duration_h']].describe()

	duration	duration_min	duration_h
count	6.699590e+05	669959.000000	669959.000000
mean	1.107950e+03	18.465831	0.307764
std	2.225544e+04	370.923950	6.182066
min	6.000000e+01	1.000000	0.016667
25%	3.440000e+02	5.733333	0.095556
50%	5.170000e+02	8.616667	0.143611
75%	7.550000e+02	12.583333	0.209722
max	1.727040e+07	287840.000000	4797.333333

max(trip_customer_df["rental_duration"])

Timedelta('199 days 22:19:00')

-- Die Mindestfahrzeit beträgt eine Minute.
-- Mittlere Reisezeit von etwa achteinhalb Minuten
-- Es gibt Ausreißer. Die maximale Reisedauer beträgt 287840 Minuten, das sind fast 200 Tage.

Deshalb überprüfen die Anzahl der Fahrten, die weniger als zwei Minuten und mehr als einen Tag dauern

Fahrten, die kürzer als zwei Minuten sind

len(trip_customer_df.loc[trip_customer_df['duration'] < 120, :]) / len(trip_customer_df) * 100

0.7934813921448924

!!! Fast ein Prozent der Fahrten unter zwei Minuten ist.

Fahrten, die länger als 24 Stunden dauern

df_trips_longer24h = trip_customer_df.loc[
        trip_customer_df['duration'] > 24*60*60 ,
        ['start_station_id', 'end_station_id', 'subscription_type', 'duration', 'duration_min', 'duration_h', 'rental_duration']
    ].sort_values("duration", ascending=False)

df_trips_longer24h

	start_station_id	end_station_id	subscription_type	duration	duration_min	duration_h	rental_duration
135291	66	62	Customer	17270400	287840.000000	4797.333333	199 days 22:19:00
416121	77	68	Customer	2137000	35616.666667	593.611111	24 days 17:37:00
630246	31	32	Subscriber	1852590	30876.500000	514.608333	21 days 10:36:00
613014	35	35	Customer	1133540	18892.333333	314.872222	13 days 02:52:00
612403	35	35	Customer	722236	12037.266667	200.621111	8 days 08:37:00
...	...	...	...	...	...	...	...
665985	3	3	Customer	86890	1448.166667	24.136111	1 days 00:09:00
608480	29	29	Customer	86668	1444.466667	24.074444	1 days 00:05:00
645219	8	8	Customer	86609	1443.483333	24.058056	1 days 00:04:00
611014	37	37	Customer	86573	1442.883333	24.048056	1 days 00:02:00
627343	32	30	Customer	86553	1442.550000	24.042500	1 days 00:03:00

296 rows × 7 columns

Wenn sich Personen ein Fahrrad länger ausleihen und bei sich stehen haben, kann niemand anderes es nutzen.

--> Daher prüfen, ob es mit dem Art des Abonnements zusammenhängt.

df_trips_longer24h["duration_d"] = df_trips_longer24h["duration_h"] / 24
df_trips_longer24h.groupby("subscription_type")["duration_d"].describe()

	count	mean	std	min	25%	50%	75%	max
subscription_type
Customer	248.0	3.292132	12.735369	1.001771	1.179682	1.914531	2.853788	199.888889
Subscriber	48.0	2.548488	3.130647	1.012998	1.125489	1.897280	2.621849	21.442014

Die Customer verhalten sich deutlich anders als die Subscriber.

fig = px.histogram(df_trips_longer24h.query("duration_d<10"),
                   x="duration_d",
                   nbins=50,
                   marginal="box",
                   color="subscription_type",
                   title='Histogramm für 1 < lange Ausleihzeiten < 10 Tage',
                   labels={'duration_d':'Leihzeitraum in Tagen', 'subscription_type': 'Kundentyp'})
fig.show()
df_trips_longer24h.query("duration_d>=10")

	start_station_id	end_station_id	subscription_type	duration	duration_min	duration_h	rental_duration	duration_d
135291	66	62	Customer	17270400	287840.000000	4797.333333	199 days 22:19:00	199.888889
416121	77	68	Customer	2137000	35616.666667	593.611111	24 days 17:37:00	24.733796
630246	31	32	Subscriber	1852590	30876.500000	514.608333	21 days 10:36:00	21.442014
613014	35	35	Customer	1133540	18892.333333	314.872222	13 days 02:52:00	13.119676

trip_customer_df.groupby("subscription_type")["duration_min"].describe()

	count	mean	std	min	25%	50%	75%	max
subscription_type
Customer	103213.0	65.862689	933.941497	1.0	10.966667	18.600000	38.816667	287840.0
Subscriber	566746.0	9.834148	57.525599	1.0	5.383333	7.933333	11.100000	30876.5

-- Die Subscriber haben eine allgemein viel kürzere Nutzungsdauer.
-- Was ist mit Dauern kleiner als eine Minute?
-- Sind die langen Nutzungen bezahlt worden? Muss dort früher nachgehackt werden?
-- Haben Kunden ein Rad dauerhaft bei sich? (Aboschwäche?)

Roundtrips

round_trips_df = trip_customer_df.query("start_station_id == end_station_id")
round_trips_df.head()

	id	duration	start_date	start_station_name	start_station_id	end_date	end_station_name	end_station_id	bike_id	subscription_type	...	Year_start	Month_start	Day_start	Day_of_week_start	Day_of_week_num_start	Time_start	rental_duration	Hour_start	duration_h	duration_min
0	4069	174	2013-08-29 09:08:00	2nd at South Park	64	2013-08-29 09:11:00	2nd at South Park	64	288	Subscriber	...	2013	8	29	Thursday	3	09:08:00	0 days 00:03:00	9	0.048333	2.900000
1	4426	982	2013-08-29 12:52:00	2nd at South Park	64	2013-08-29 13:08:00	2nd at South Park	64	593	Subscriber	...	2013	8	29	Thursday	3	12:52:00	0 days 00:16:00	12	0.272778	16.366667
2	4557	130	2013-08-29 13:57:00	2nd at South Park	64	2013-08-29 13:59:00	2nd at South Park	64	371	Subscriber	...	2013	8	29	Thursday	3	13:57:00	0 days 00:02:00	13	0.036111	2.166667
3	4564	1693	2013-08-29 14:04:00	2nd at South Park	64	2013-08-29 14:32:00	2nd at South Park	64	272	Subscriber	...	2013	8	29	Thursday	3	14:04:00	0 days 00:28:00	14	0.470278	28.216667
4	4566	1699	2013-08-29 14:04:00	2nd at South Park	64	2013-08-29 14:33:00	2nd at South Park	64	371	Subscriber	...	2013	8	29	Thursday	3	14:04:00	0 days 00:29:00	14	0.471944	28.316667

5 rows × 27 columns

round_trips_df.groupby(["subscription_type"])["duration_min"].describe()

	count	mean	std	min	25%	50%	75%	max
subscription_type
Customer	16443.0	143.545195	371.252668	1.0	26.133333	67.666667	149.225000	18892.333333
Subscriber	7538.0	23.963072	121.404286	1.0	3.700000	12.041667	21.733333	4831.216667

fig = px.histogram(
    round_trips_df.query('duration_min<(8*60)'),
    x="duration_min",
    nbins=50,
    marginal="box",
    color="subscription_type",
    title='Histogramm für Roundtrips < 6h')
fig.show()

round_trips_df.query('duration_min>=(8*60)').groupby("subscription_type")["duration_min"].describe()

	count	mean	std	min	25%	50%	75%	max
subscription_type
Customer	632.0	1297.459599	1399.084513	480.450000	588.991667	1053.450000	1376.470833	18892.333333
Subscriber	40.0	1212.813750	1102.710294	483.366667	544.162500	815.216667	1276.075000	4831.216667

-- Die Customer scheinen Roundtrips eher lange und häufiger zu machen in Form von Ausflügen. Die Subscriber machen eher kürzere Fahrten (Besorgungen?)

Intercity-Fahrten

trip_customer_df.loc[trip_customer_df['city_start'] != trip_customer_df['city_end'], ['bike_id', 'rental_duration', 'lat_start', 'long_start', 'city_start', 'lat_end', 'long_end', 'city_end']]

	bike_id	rental_duration	lat_start	long_start	city_start	lat_end	long_end	city_end
41365	180	0 days 06:17:00	37.486076	-122.232086	Redwood City	37.776615	-122.395264	San Francisco
41366	64	0 days 06:13:00	37.486076	-122.232086	Redwood City	37.776615	-122.395264	San Francisco
42491	686	0 days 05:30:00	37.444523	-122.163094	Palo Alto	37.776615	-122.395264	San Francisco
42492	168	0 days 05:29:00	37.444523	-122.163094	Palo Alto	37.776615	-122.395264	San Francisco
42493	240	0 days 05:08:00	37.444523	-122.163094	Palo Alto	37.776615	-122.395264	San Francisco
...	...	...	...	...	...	...	...	...
662943	98	0 days 01:41:00	37.394360	-122.076714	Mountain View	37.348743	-121.894714	San Jose
664011	41	0 days 02:57:00	37.400242	-122.099075	Mountain View	37.348743	-121.894714	San Jose
664012	230	0 days 02:54:00	37.400242	-122.099075	Mountain View	37.348743	-121.894714	San Jose
666920	13	0 days 03:05:00	37.406940	-122.106758	Mountain View	37.330696	-121.888977	San Jose
668397	102	0 days 07:37:00	37.390278	-122.066551	Mountain View	37.332691	-121.900085	San Jose

1042 rows × 8 columns

-- Die Daten enthält Informationen über 1.042 Intercity-Fahrten

Durchschnittliche Fahrtdauer je nach Art des Abonnements und Wochentag

customer_trips_df.loc[:, ['duration', 'duration_min', 'duration_h']].describe()

	duration	duration_min	duration_h
count	1.032130e+05	103213.000000	103213.000000
mean	3.951761e+03	65.862689	1.097711
std	5.603649e+04	933.941497	15.565692
min	6.000000e+01	1.000000	0.016667
25%	6.580000e+02	10.966667	0.182778
50%	1.116000e+03	18.600000	0.310000
75%	2.329000e+03	38.816667	0.646944
max	1.727040e+07	287840.000000	4797.333333

subscriber_trips_df.loc[:, ['duration', 'duration_min', 'duration_h']].describe()

	duration	duration_min	duration_h
count	5.667460e+05	566746.000000	566746.000000
mean	5.900489e+02	9.834148	0.163902
std	3.451536e+03	57.525599	0.958760
min	6.000000e+01	1.000000	0.016667
25%	3.230000e+02	5.383333	0.089722
50%	4.760000e+02	7.933333	0.132222
75%	6.660000e+02	11.100000	0.185000
max	1.852590e+06	30876.500000	514.608333

def percentile(n):
    def percentile_(x):
        return np.percentile(x, n)
    percentile_.__name__ = 'percentile_%s' % n
    return percentile_

duration_day_df = trip_customer_df.groupby(
    ['subscription_type', 'Day_of_week_start', 'Day_of_week_num_start'], as_index=False, observed=True).agg(
    duration_median = ('duration_min', np.median),
    #duration_q25 = ('duration_min', percentile(25)),
   # duration_q75 = ('duration_min', percentile(75))
    
).sort_values('Day_of_week_num_start')

duration_day_df.head()

	subscription_type	Day_of_week_start	Day_of_week_num_start	duration_median
1	Subscriber	Monday	0	7.900000
8	Customer	Monday	0	17.483333
5	Subscriber	Tuesday	1	7.850000
12	Customer	Tuesday	1	15.783333
6	Subscriber	Wednesday	2	7.933333

fig = px.line(duration_day_df, x="Day_of_week_start", y="duration_median",
              title='Durchschnittliche Fahrtdauer je nach Art des Abonnements und Wochentag',
              color = 'subscription_type',
              labels={'duration_median':'mediane Fahrtdauer min', 'Day_of_week_start':'Wochentag', 'subscription_type' : 'Abonnementstyp'})


fig.update_traces(mode="markers+lines")

fig.show()

An welchen Stationen es Probleme gibt?

print(round(len(no_bikes_or_docks_df) / 71984434 * 100, 2), '%')
print(round(len(critical_bikes_or_docks_df) / 71984434 * 100, 2), '%')

1.19 %
2.44 %

-- Für über 1% der Zeit sind die Stationen nicht voll nutzbar.
-- Und zu etwa 2.5% der Zeit sind an den Stationen nur noch eine kritische Anzahl an Docks oder Fahrrädern verfügbar.

critical_bikes_or_docks_df['reason'] = 'critical'

no_bikes_or_docks_df['reason'] = no_bikes_or_docks_df.bikes_available.apply(lambda x: "bikes" if x == 0  else "docks")

print(no_bikes_or_docks_df.columns)
print(critical_bikes_or_docks_df.columns)

Index(['station_id', 'bikes_available', 'docks_available', 'time', 'reason'], dtype='object')
Index(['station_id', 'bikes_available', 'docks_available', 'time', 'reason'], dtype='object')

no_bikes_or_docks_gr = no_bikes_or_docks_df.groupby(
    ['station_id', 'reason'], as_index = False).size().sort_values('station_id')

no_bikes_or_docks_gr = no_bikes_or_docks_gr.merge(station_info_df, left_on='station_id', right_on='id')
no_bikes_or_docks_gr['time_problem_h'] = round(no_bikes_or_docks_gr['size'] / 60)

no_bikes_or_docks_gr['time_problem_h'] =no_bikes_or_docks_gr['time_problem_h'].fillna(0)

no_bikes_or_docks_gr.loc[no_bikes_or_docks_gr['station_id'] == 31, :]

	station_id	reason	size	id	name	lat	long	dock_count	city	installation_date	time_problem_h
41	31	bikes	4036	31	San Antonio Shopping Center	37.400443	-122.108338	15	Mountain View	2013-12-31	67.0
42	31	docks	1035	31	San Antonio Shopping Center	37.400443	-122.108338	15	Mountain View	2013-12-31	17.0

fig = px.bar(no_bikes_or_docks_gr, x='station_id', y='size', title="Volle und leere Stationen",
             color = 'reason', labels={'size':'Dauer in Minuten'},
             hover_data =["city", 'station_id', 'size'])

fig.show()

fig = px.scatter_mapbox(no_bikes_or_docks_gr,
                        lat="lat",
                        lon="long",
                        mapbox_style="carto-positron",
                        color="reason",
                        size="time_problem_h",
                        color_discrete_sequence=px.colors.qualitative.G10,
                        size_max=20,
                       hover_data =['station_id'])
fig.show()

station_use_df = trip_customer_df.groupby('start_station_id', as_index=False).size()
station_use_df.rename(columns={'size' : 'use_num'}, inplace=True)
len(station_use_df)

70

station_problem_df = station_problem_df.merge(station_use_df, left_on='station_id', right_on='start_station_id')
station_problem_df.drop(columns=['start_station_id'], inplace=True)
station_problem_df

	station_id	no_bikes_min	no_docks_min	use_num
0	2	1340	466	9558
1	3	1066	4817	1594
2	4	10618	6614	3861
3	5	1321	0	1257
4	6	3564	3633	2917
...	...	...	...	...
65	77	4436	3129	24172
66	80	2105	5	863
67	82	17824	3738	10310
68	83	808	0	341
69	84	3642	1744	1747

70 rows × 4 columns

station_problem_df = station_problem_df.merge(
    station_info_df.loc[:, ['id', 'lat', 'long', 'name', 'city', 'dock_count']],
    left_on='station_id',
    right_on='id')

station_problem_df.drop(columns=['id'], inplace=True)
station_problem_df

	station_id	no_bikes_min	no_docks_min	use_num	lat	long	name	city	dock_count
0	2	1340	466	9558	37.329732	-121.901782	San Jose Diridon Caltrain Station	San Jose	27
1	3	1066	4817	1594	37.330698	-121.888979	San Jose Civic Center	San Jose	15
2	4	10618	6614	3861	37.333988	-121.894902	Santa Clara at Almaden	San Jose	11
3	5	1321	0	1257	37.331415	-121.893200	Adobe on Almaden	San Jose	19
4	6	3564	3633	2917	37.336721	-121.894074	San Pedro Square	San Jose	15
...	...	...	...	...	...	...	...	...	...
65	77	4436	3129	24172	37.789625	-122.400811	Market at Sansome	San Francisco	27
66	80	2105	5	863	37.352601	-121.905733	Santa Clara County Civic Center	San Jose	15
67	82	17824	3738	10310	37.798541	-122.400862	Broadway St at Battery St	San Francisco	15
68	83	808	0	341	37.491269	-122.236234	Mezes Park	Redwood City	15
69	84	3642	1744	1747	37.342725	-121.895617	Ryland Park	San Jose	15

70 rows × 9 columns

station_problem_df['no_bikes_h'] = station_problem_df['no_bikes_min'] / 60
station_problem_df['no_docks_h'] = station_problem_df['no_docks_min'] / 60

fig = px.scatter(station_problem_df, x="no_bikes_h", y="no_docks_h", color='city',
                 hover_data=["station_id", "name", 'dock_count'],
                 size="use_num",
                 labels={'no_bikes_h':'Ohne Fahrräder (ST)', 'no_docks_h': 'Ohne freie Dockstationen (ST)', 'city':'Stadt'},
                 title='')

fig.show()

Animation: San Francisco 16 April 2014 (Mittwoch)

df_sf_16_april_2014.sort_values(by=['station_id', 'time_of_day'], ascending=True)

	station_id	bikes_available	docks_available	time	lat	long	time_of_day	hour
0	39	14	5	2014-04-16 00:00:03	37.783871	-122.408433	00:00:03	0
1	39	14	5	2014-04-16 00:01:03	37.783871	-122.408433	00:01:03	0
2	39	14	5	2014-04-16 00:02:03	37.783871	-122.408433	00:02:03	0
3	39	14	5	2014-04-16 00:03:02	37.783871	-122.408433	00:03:02	0
4	39	14	5	2014-04-16 00:04:02	37.783871	-122.408433	00:04:02	0
...	...	...	...	...	...	...	...	...
50395	82	6	9	2014-04-16 23:55:03	37.798541	-122.400862	23:55:03	23
50396	82	6	9	2014-04-16 23:56:02	37.798541	-122.400862	23:56:02	23
50397	82	6	9	2014-04-16 23:57:03	37.798541	-122.400862	23:57:03	23
50398	82	6	9	2014-04-16 23:58:03	37.798541	-122.400862	23:58:03	23
50399	82	6	9	2014-04-16 23:59:03	37.798541	-122.400862	23:59:03	23

50400 rows × 8 columns

type(df_sf_16_april_2014['time_of_day'][0])

datetime.time

df_sf_16_april_2014['time_of_day_str'] = df_sf_16_april_2014['time_of_day'].apply(lambda x: str(x))

fig = px.scatter_mapbox(
           df_sf_16_april_2014.loc[::8,:],
           lat="lat", lon="long", animation_frame="time_of_day_str", animation_group="station_id",
           size="bikes_available",
           color='bikes_available',
           mapbox_style="carto-positron", range_color=(0,28),
           center={"lat": 37.785, "lon": -122.405}, zoom=11.8, size_max=25,
           labels={'time_of_day_str':'Zeit', 'bikes_available': 'Fahrräder verfügbar'},
           title='San Francisco 16.04.2014'
)

fig.show()

Station 'San Francisco Caltrain (Townsend at 4th)' (ID = 70)

station_70_df

	station_id	bikes_available	docks_available	time
0	70	0	19	2013/08/29 12:06:01
1	70	0	19	2013/08/29 12:07:01
2	70	0	19	2013/08/29 12:08:01
3	70	0	19	2013/08/29 12:09:01
4	70	0	19	2013/08/29 12:10:01
...	...	...	...	...
1047135	70	15	4	2015-08-31 23:55:02
1047136	70	15	4	2015-08-31 23:56:01
1047137	70	15	4	2015-08-31 23:57:02
1047138	70	15	4	2015-08-31 23:58:02
1047139	70	15	4	2015-08-31 23:59:02

1047140 rows × 4 columns

Verfügbarkeit des Services

st_70_no_bikes = station_70_df.loc[station_70_df.bikes_available == 0].shape[0]
st_70_no_docks = station_70_df.loc[station_70_df.docks_available == 0].shape[0]
st_70_critical_bikes = station_70_df.loc[station_70_df.bikes_available == 1].shape[0]
st_70_critical_docks = station_70_df.loc[station_70_df.docks_available == 1].shape[0]

st_70_used_num = station_70_df.shape[0] - st_70_no_bikes - st_70_no_docks - st_70_critical_bikes - st_70_critical_docks

d = {
        'param': ['no_bikes', 'no_docks', 'critical_bikes', 'critical_docks', 'ok'],
         'val': [st_70_no_bikes, st_70_no_docks, st_70_critical_bikes, st_70_critical_docks, st_70_used_num]
}
station_70_sum_df = pd.DataFrame(data=d)
station_70_sum_df

	param	val
0	no_bikes	21726
1	no_docks	19422
2	critical_bikes	22596
3	critical_docks	49381
4	ok	934015

fig = px.pie(station_70_sum_df, values = 'val', names='param', title= 'Verfügbarkeit des Services')
fig.show()

Durchschnittliche Anzahl der während der Woche verfügbaren Fahrräder

station_70_df.loc[:, 'time'] = pd.to_datetime(station_70_df.loc[:, 'time'])

station_70_df['Day_of_week'] = station_70_df.loc[:, 'time'].dt.day_name() 
station_70_df['Day_of_week_num'] = station_70_df.loc[:, 'time'].dt.day_of_week
station_70_df['time_h'] = station_70_df.loc[:, 'time'].dt.hour
station_70_df

	station_id	bikes_available	docks_available	time	Day_of_week	Day_of_week_num	time_h
0	70	0	19	2013-08-29 12:06:01	Thursday	3	12
1	70	0	19	2013-08-29 12:07:01	Thursday	3	12
2	70	0	19	2013-08-29 12:08:01	Thursday	3	12
3	70	0	19	2013-08-29 12:09:01	Thursday	3	12
4	70	0	19	2013-08-29 12:10:01	Thursday	3	12
...	...	...	...	...	...	...	...
1047135	70	15	4	2015-08-31 23:55:02	Monday	0	23
1047136	70	15	4	2015-08-31 23:56:01	Monday	0	23
1047137	70	15	4	2015-08-31 23:57:02	Monday	0	23
1047138	70	15	4	2015-08-31 23:58:02	Monday	0	23
1047139	70	15	4	2015-08-31 23:59:02	Monday	0	23

1047140 rows × 7 columns

station_70_df_gr = station_70_df.groupby(['station_id', 'Day_of_week', 'time_h'], as_index=False).agg(
    bikes_available_mean = ('bikes_available', np.mean),
    docks_available_mean = ('docks_available', np.mean))
station_70_df_gr

	station_id	Day_of_week	time_h	bikes_available_mean	docks_available_mean
0	70	Friday	0	14.501036	4.489399
1	70	Friday	1	14.524098	4.466326
2	70	Friday	2	14.525052	4.465405
3	70	Friday	3	14.510987	4.479459
4	70	Friday	4	14.524908	4.465542
...	...	...	...	...	...
163	70	Wednesday	19	12.299020	6.570259
164	70	Wednesday	20	12.455159	6.459820
165	70	Wednesday	21	13.679801	5.307977
166	70	Wednesday	22	13.903775	5.087711
167	70	Wednesday	23	13.746343	5.244012

168 rows × 5 columns

fig = px.line(station_70_df_gr,
                     x='time_h',
                     y='bikes_available_mean',
                     color='Day_of_week',
                     labels={'bikes_available_mean':'Anzahl verfügbarer Fahrräder',
                            'Day_of_week':'Wochentag',
                            'time_h': 'Zeit'},
                     title='Durchschnittliche Anzahl der Fahrräder, die an der Station 70 während der Woche zur Verfügung stehen',
                     hover_name='Day_of_week')
fig.show()

fig = px.line(station_70_df_gr,
                     x='time_h',
                     y='docks_available_mean',
                     color='Day_of_week',
                     labels={'docks_available_mean':'Anzahl verfügbarer Dockstationen',
                             'size':'Counts of Available Bikes-Numbers',
                            'Day_of_week':'Wochentag',
                            'time_h': 'Zeit'},
                     title='Durchschnittliche Anzahl der Dockstationen, die an der Station 70 während der Woche zur Verfügung stehen',
                     hover_name='Day_of_week')
fig.show()

Wie die Anzahl der verfügbaren Fahrräder im Laufe des Tages variiert?

station_status_df.head()

	station_id	avg_bikes_available	avg_docks_available	day_of_week	h
0	2	13.675880	13.295195	0	00
1	2	13.610708	13.360264	0	01
2	2	13.417803	13.552636	0	02
3	2	13.505408	13.465537	0	03
4	2	13.496455	13.474541	0	04

station_status_df['day_of_week'] = station_status_df['day_of_week'].astype(int)
station_status_df['h'] = station_status_df['h'].astype(int)

# station_status_df
station_status_df["day_type"] = station_status_df.day_of_week.apply(lambda x: "Working" if x < 5  else "Weekend")

station_status_df = station_status_df.merge(
    station_info_df.loc[:, ['id', 'city', 'lat', 'long', 'name']],
    left_on='station_id', right_on='id')

station_status_df

	station_id	avg_bikes_available	avg_docks_available	day_of_week	h	day_type	id	city	lat	long	name
0	2	13.675880	13.295195	0	0	Working	2	San Jose	37.329732	-121.901782	San Jose Diridon Caltrain Station
1	2	13.610708	13.360264	0	1	Working	2	San Jose	37.329732	-121.901782	San Jose Diridon Caltrain Station
2	2	13.417803	13.552636	0	2	Working	2	San Jose	37.329732	-121.901782	San Jose Diridon Caltrain Station
3	2	13.505408	13.465537	0	3	Working	2	San Jose	37.329732	-121.901782	San Jose Diridon Caltrain Station
4	2	13.496455	13.474541	0	4	Working	2	San Jose	37.329732	-121.901782	San Jose Diridon Caltrain Station
...	...	...	...	...	...	...	...	...	...	...	...
11755	84	7.027626	7.958676	6	19	Weekend	84	San Jose	37.342725	-121.895617	Ryland Park
11756	84	7.094770	7.891528	6	20	Weekend	84	San Jose	37.342725	-121.895617	Ryland Park
11757	84	7.171005	7.815297	6	21	Weekend	84	San Jose	37.342725	-121.895617	Ryland Park
11758	84	7.242466	7.743836	6	22	Weekend	84	San Jose	37.342725	-121.895617	Ryland Park
11759	84	7.290183	7.696119	6	23	Weekend	84	San Jose	37.342725	-121.895617	Ryland Park

11760 rows × 11 columns

station_status_df_gr = station_status_df.groupby(
    ['station_id', 'h', 'day_type', 'city', 'lat', 'long'],
    as_index=False).agg(
                        avg_bikes_available = ('avg_bikes_available', np.mean),
                        avg_docks_available = ('avg_docks_available', np.mean))

max(station_status_df.avg_bikes_available)

18.846276257161044

fig_station_dev = px.scatter_mapbox(data_frame=station_status_df_gr.loc[station_status_df_gr.day_type == 'Working', :],
                                    title="Nutzer Zeiten während Wochentage", 
                                    lat='lat', 
                                    lon='long',
                                    size="avg_bikes_available",
                                    color="avg_bikes_available",
                                    color_continuous_scale="greens",
                                    range_color=(0,19),
                                    size_max=25,
                                    animation_frame="h", 
                                    mapbox_style="carto-positron",
#                                     width=1000, 
                                    height=850)

fig_station_dev.show()